Free PPC

Toolkit

You have an hour with an inherited account before the kickoff call and the client wants to hear what is broken. This is the structure I run, in this order, so the audit ends with three priced fixes instead of twenty observations.

The framework

Walk these in order. Each gate filters out what is irrelevant before the next dimension is read.

Diagnostic tree by symptom

What this framework solves

Most Google Ads audits become checklists that grade every dimension and surface twenty problems of equal weight. The result is a report that gets filed, not implemented. This framework replaces breadth with depth: find the one structural cause behind the noise, quantify it in dollars, and sequence the fix order so the audit changes the account.

The core idea

Every account has one root cause, not twenty problems. Three ad groups with high CPA, poor Quality Scores, and low impression share usually trace back to a single fragmented campaign structure or a broken tracking setup, and the rest are downstream symptoms. Averages hide breaks: a campaign averaging $45 CPA can contain ad groups at $25 and $90, so every metric must be disaggregated before it earns a verdict. Findings expressed in percentages get ignored; findings expressed in dollars get implemented. And before any of this, check whether the problem is even fixable inside Google Ads, because a conversion collapse from a broken landing page is not a campaign problem.

How to apply it

1. Validate tracking before anything else. A 30 to 50 percent tracking discrepancy makes every metric below it unreliable. If tracking is broken, the audit has one finding and it is “fix tracking.”
2. Classify the account’s maturity stage. Under 30 days: structural review only, no performance diagnosis. Mature plateau: demand ceiling analysis, not more keyword optimization. Turnaround: triage to the single biggest waste source plus three actions and a 30-day checkpoint. The stage determines which analytical tools are valid.
3. Walk the diagnostic trees by symptom. Sudden conversion drop traces to tracking, then page changes, then disapprovals, then competition, before any campaign-level suspicion. CPA above target across all campaigns asks first whether the target is realistic. Strong ROAS with low impression share asks whether the loss is budget or rank. PMax appearing as top performer triggers a brand-cannibalization check.
4. Disaggregate every metric before declaring it healthy. Drill from account to campaign to ad group to keyword to search term, and from blended to brand vs. non-brand. Aggregated numbers always hide the break.
5. Sequence findings as waste-first, then winners, then growth. Stop the bleeding, feed the proven performers, then test expansion. An audit that recommends scaling before quantifying waste is asking the client to fund waste at scale.
6. Quantify in dollars, not percentages. “$730 per month wasted on irrelevant search terms” is a finding. “CPA is 15 percent above target” is an observation. Use a $500 per month minimum impact threshold for dedicated finding cards in well-managed accounts.
7. Calibrate confidence to volume. 50+ conversions: directive recommendations. 20 to 50: caveats. 10 to 20: directional only. Below 10: ranges, not point estimates. Treating eight conversions like eighty produces bad recommendations.
8. Deliver the fix, not just the diagnosis. “Add negatives” without the keyword list is incomplete. “Consolidate campaigns” without the proposed structure is incomplete. The deliverable is the action.

When to use it

Use this on any inherited or under-performing account, before any optimization work, and on any account where the brief is “tell me what is wrong.” It applies across verticals with calibration: ecommerce leads with feed health and ROAS, lead gen leads with offline conversion gaps and effective CAC, B2B SaaS works at account-level aggregation because campaign-level volume is too thin, local services starts with demand ceiling math, and high-value verticals lower the waste threshold because $150 CPCs make small leaks expensive fast. It does not apply when the underlying offer or unit economics cannot support paid acquisition at all. No audit fix rescues an account that should not be advertising.